Re: Looking for good Ada95 book

[dewar@merv.cs.nyu.edu (Robert Dewar)]

Michael says

"> I was not asked what my opinion was before forming this imaginary
> consensus, and I feel that we should reject all nonsense standards
> such as capitalization that take time away from making a program
> semantically correct (solving the problem correctly) and making
> it efficient."

(at least I hope it was Michael, sorry if I am confused).

In fact consistency of style is critical to producting reliable
maintainable programs. Sure, there are always some people who say
that they don't want to be constrained by such silly things as
petty consistency. I consider this viewpoint to be a menace since
it leads to notions of idiosyncratic personal style and code
ownership, which to me are anathema to realiable large scale
programming.

I think Michael's specific recommended style (lower case for everything)
does NOT lead to the most readable code, and I think it is a big
improvement that we are indeed seeing something like a general consensus
in coding style (I react here to the huge volume of Ada 95 code we receive
at report@gnat.com from thousands of users of GNAt, nearly all of whom adhere
to the lower case keywords, Mixed_Case_Identifier style. Not all, but nearly
all.

I know that one can have tools that automatically translate from one style
to another, but even if such tools work perfectly (they cannot in practice,
since naming often has semantic contnts, e.g. NOSC_Policy, not Nosc_Policy
may be preferred), they do not promote common interchange of code. It is
far better if, as in the C world, people converge on a single style that
everyone uses.

I think this convergence is much more important than what particular style
is chosen. As Tarski use to say, explaining why stanards are important,
it really doesn't matter if you drive on the left or the right, but it is
quite importnat that everyone agree.

I used to prefer the Ada-83 RM SHOUTING_IDENTIFIER style, since it is wht
I was used to, but after a painful transition, I now can't stand that
style any more. That's the real point, people get used to almost any
convention, so they might as well get used to a common comvention.

That is why Mike's book is disturbing, it creates a generation of Ada
programmers who have got used to a seriously non-standard style. Virtually
NO code that we see at report@gnat.com uses upper case keywords like BEGIN
and END. The only other use we see is SHOUTING_IDENTIFIERS, Ada 83 style, and
that is certainly disappearing fast.